Accurate identification of samples within a laboratory is vital to correctly linking a diagnosis to a patient. The correlation of tissue sample sections, such as thin tissue sections on a microscope slide, with the tissue sample support, such as a paraffin-embedded tissue block from which the sections were obtained, is critical to accurate analysis and diagnosis.
Typically, the workflow within a histology laboratory is as follows:
A biopsy sample is delivered in a container by courier from a doctor's surgery. The sample container is given an accession ID relevant to the laboratory, and is then passed on to the cut-up (grossing) area. Here, the sample is removed from the container, and cut up to excise pieces of interest. The pieces of interest are placed into one or more tissue cassettes, which are uniquely labelled to link them to the accession ID. Any remaining specimen tissue is returned to the initial container for storage.
The tissue cassettes are batched into a collection basket, and when sufficient are available they are placed into a tissue processor where the tissue pieces are infiltrated with wax to support the tissues and allow cutting of thin sections (eg 2-10 microns) at the microtome. The collection basket is transferred to the embedding work-station. Here, working on a single cassette at a time to minimise errors, tissue pieces are removed from the cassettes and placed into a mould. Molten wax is poured into the mould, and the tissue cassette base, which carries the identification information, is placed on the top of the mould. When the mould is cooled, a tissue block is produced in which the paraffin holds together the cassette base and the embedded pieces, allowing for identification of the tissues embedded in the paraffin block.
The embedded tissue block is passed to the microtome operator, who places identification information relating to the tissue block onto one or more microscope slides, as necessary, and then places the block into the microtome chuck. Thin sections (eg 2-10 microns) are cut from the block and transferred onto the identified microscope slides, traditionally by floating the paraffin sections onto a water-bath and then scooping them up with the microscope slide.
Slides are then batched into a drying rack, transferred into a staining rack and placed into a staining mechanism, where features of the tissues are stained in various colours. Once stained, the slides have a thin glass coverslip applied to protect the tissue and improve the optical properties of the slide for viewing in a microscope. At this point, the slides are ready to pass to a pathologist for diagnosis.
However, due to the possibility of identification errors in the above process, two further time-consuming steps are required. Firstly, a machine printed label is applied. It is usual that the initial identification markings placed on the slide by the microtome operator are handwritten in pencil. Pencil markings are used because they are resistant to the chemicals used in slide staining. Unfortunately, hand-written pencil markings are often hard to read and are not smudgeproof, and so after staining it is common to apply a machine-printed label to the slide to provide an easily read and smudgeproof identification. The application of a label at this step allows for human error in identifying the sample.
Secondly, slides and tissue blocks are collated to reduce identification errors. In this step, a human operator examines each and every slide and tissue block to verify that the tissue on the slide does indeed come from the tissue block that the label identifies it as coming from, and that the tissue in the block is of the correct type and size as expected. This collation step is labour intensive and provides significant delays in diagnosis. Finally, after collation, tissue blocks are filed and slides can be distributed to pathologists for diagnosis.
In general, errors may arise anywhere where a sample is moved from one container to another, where sections are taken from the sample, or when a sample is re-labelled. As can be seen, in the above processing workflow five separate identification steps, and a checking step are required:                1. Tissue samples arrive in containers and an accession ID is attached to the container.        2. The tissue is removed from the container, cut up and the pieces are placed into cassettes—the cassette IDs are required to identify not only the sample from which the pieces come, but also the part of the sample to which they relate.        3. Processed tissues pieces are embedded into a paraffin tissue block, and the cassette ID is transferred to the tissue block by attaching the base of the cassette to the paraffin mould.        4. At the microtome, sections are cut from the samples and are placed onto microscope slides. The slides need to be marked in such a way as to identify the sample from which they came, usually by copying the cassette identifying marks onto the slide in pencil by hand. Multiple cassettes may be used per sample, and multiple slides can be required per cassette.        5. After staining, the hand-written slides are re-labelled with machine-printed labels.        6. A final collation step is required to ensure that the stained tissues on the slides, and the embedded tissues in the tissue block, match with the accessioned tissue sample.        
Identification errors may occur at any one of these steps.
Furthermore, a consequence of the error-prone nature of current slide identification practices is that slide identification is not guaranteed until after staining—pre-stained slides contain sections that are so thin as to be optically transparent, and so features required for identification cannot yet be seen. Therefore collation is necessarily done after staining.
Automation of processes in a histology laboratory is seen as beneficial to improving quality and reducing turn around time, and can be assisted by automatic identification of the slide, eg using barcoding. Using current identification practices however, it is not possible to ensure error-free processing of slides based on machine-readable identification. Identification errors can only be picked up after staining has occurred, and correction of these errors requires time-consuming and costly re-testing, and involves the use of additional sample tissue. In many cases, this additional sample tissue may not be available.
A number of practices have evolved over the years to address the issue of cross-referencing. These practices include use of paper-based systems to identify cassette contents at the microtome, wherein a file of papers detailing the samples are passed along with the cassette and slide. These paper-based systems are generally not used, as transfer of paper records is clumsy, requires duplicate copies, and requires complex choices about workflow, such as whether the paper follows the cassettes, slides, both or neither. Paper-based systems also require additional desk space as well as the potential problem of paper records getting covered in paraffin and/or chemicals.
A major source of cross-referencing errors in the present process is slide identification. Currently, slide labels are predominantly hand-written, with the microtome operator transcribing the cassette identification information by hand to the slide. This is a time-intensive task with significant potential for human error. Handwriting requires handling of slides, which is especially undesirable as it may transfer skin cells to the slide surface. Additionally, handwriting clearly does not allow for machine-readable information to be created, and is thus not amenable to automatic identification of the slide.
A number of variations on the above standard workflow have evolved to address some of the shortcomings of hand-written labels. One such method, with a similar workflow to handwriting of slides, is to print sticky labels (using, for example, a thermal transfer printer) and hand-apply these labels to the slides. The labels can either be printed at the time of placing the sample on the slide or, alternatively, labels can be printed in batches beforehand. Application of labels is a time-consuming process that still requires handling of the slides, as no automatic label applicator yet exists that is used at a microtome station. Additionally, hand-application of labels is prone to the same human-error as hand-writing of slides, in that the human operator can mislabel slides.
Printers have been designed that are capable of printing directly onto slides using chemical resistant, smudgeproof inks. Such printers are expensive and bulky, and are not designed for, or suitable for, use at the microtome station. As a lab cannot generally afford more than one, they are used for printing large batches of slides. Similarly, label printers are capable of printing large batches of labels.
Another alternative workflow to that outlined above for handwriting is to use a printer to pre-print batches of slides or labels, and distribute these slides, labels or pre-labelled slides together with the tissue blocks to the microtome operators. While this batch pre-printing process is common, it requires an additional time consuming and error-prone process of sorting marked slides or labels and allocating them, together with the tissue blocks, to the microtome operator. Furthermore, when using batch-pre-printing processes it is impossible to guarantee that the correct number of slides/labels are produced. This is because changes to the number of slides required can be made in the grossing (cut-up), embedding, and sectioning (microtome) stages of histology processing. This leads to both wastage of slides and labels that have been unnecessarily printed, and also to wasted time where the microtome operator must either hand-write additional slides, or print further slides on the batch printer.
It can be seen that there are problems with all of the above workflows. To date, no complete automated solution to the problem of cross-referencing of tissue holders and tissue supports has been developed, and no known process provides an automated identification prior to staining.
The present invention is directed to overcoming or at least alleviating the problems associated with the prior art.